Nonexercise Models for Estimating VO sub(2max) with Waist Girth, Percent Fat, or BMI

Previously published nonexercise models using either percent fat or body mass index (BMI) as body composition measures provided valid estimates of VO sub(2max). This study was conducted to investigate the use of waist girth (WG) as a body composition surrogate in the nonexercise models and to compare the accuracy of nonexercise models that include WG, %fat, or BMI. A total of 2417 men and 384 women were measured for VO sub(2max) by indirect calorimetry (RER > 1.1); age (yr); gender by M = 1, W = 0; self-report activity habit by the 11-point (0-10) NASA physical activity status scale (PASS); WG at the apex of the umbilicus; %fat by skinfolds; and BMI by weight (kg) divided by height squared (m super(2)). Three models were developed by multiple regression to estimate VO sub(2max) from age, gender, PASS, and either WG (R = 0.81, standard error of estimate (SEE) = 4.80 mL times kg super(-1) times min super(-1)), %fat (R = 0.82, SEE = 4.72 mL times kg super(-1) times min super(-1)), or BMI (R = 0.80, SEE = 4.90 mL times kg super(-1) times min super(-1)). Cross-validation by the PRESS technique confirmed these statistics. Accuracy of the models for predicting VO sub(2max) of subsamples was supported by constant errors (CE) < 1 mL times kg super(-1) times min super(-1) for subgroups of gender, age, PASS, and VO sub(2max) between 30 and 50 mL times kg super(-1) times min super(-1) (70% of the sample). CE were > 1 mL times kg super(-1) times min super(-1) for VO sub(2max) < 30 and > 50 mL times kg super(-1) times min super(-1). Waist girth is an acceptable surrogate for body composition in the nonexercise models. All models were similar in accuracy and valid for estimating VO sub(2max) of most adults, but with reduced accuracy at the extremes of fitness (VO sub(2max) < 30 and > 50 mL times kg super(-1) times min super(-1)).